Figure 11 from Text2EL+: Expert Guided Event Log Enrichment Using Unstructured Text | Semantic Scholar (2024)

Text2EL is introduced, a two-phase event log enrichment approach based on unstructured text that applies techniques from natural language processing, sentence embeddings, and contextual and expression validation before enriching the event log.

This paper combines the analysis of textual attribute values, based on a state-of-the-art language model, with a novel attribute classification technique, and extracts information about up to eight semantic roles per event through so-called semantic role labeling of event data.

This paper proposes an original framework to mine processes from CRM data, by leveraging the unstructured part of the data, using Latent Dirichlet Allocation (LDA), an unsupervised machine learning technique, to automatically detect and assign labels to activities.

The Rapid Business Process Discovery (R-BPD) framework and prototype tool is proposed that can query heterogeneous information resources and rapidly construct prototype models to be incrementally adjusted to correctness by an analyst, which will increase analyst productivity by significant orders of magnitude over traditional approaches.

A novel unsupervised technique TextProcessMiner is created that discovers process instance models from unstructured, text-based process traces using natural language processing with a focus on the verb semantics and emerged as domain independent.

An approach that utilizes NLI to derive topics and process activities from customer service conversations and that represents them in a standardized XES event log is developed and shows that NLI helps construct event logs of high accuracy for process mining purposes.

This work proposes to extract topical phrases (keyphrases) from the unstructured data using an un-supervised graph-based approach and captures underlying process interactions which allows to understand in detail the process realities and identify opportunities for improvement.

An overview of the current state-of-the-art in text-to-model mining from natural language text is given, which covers both advantages and disadvantages of current techniques.

This paper presents the first automated approach for the extraction of declarative process models from natural language techniques that identify activities and their inter-relations from textual constraint descriptions and provides automated support for an otherwise tedious and complex manual endeavor.

This paper proposes a gamified crowdsourcing approach to the detection and repair of problematic activity labels, namely those with identical semantics but different syntax, which shows promising results in terms of quality improvements achieved.

...

...